Progressive Internet Explorer | Hacker News Reader

Progressive Internet Explorer

github.com

64 pointstilt31 comments

I've used PIE on a handful of projects. Performance only suffers if you're attaching it to a ton of elements. Using it here and there doesn't cost too much.

I have to question if we should be supporting IE6 at all with these endeavours. I see it as encouraging people to continue using busted tools that are clear and present dangers to the web environment.

I get the counter-argument, but on the whole, I'd rather see IE6,7 and 8 just die. If corporations are too risk-averse to get rid of a clearly-broken tool, why should we help them out?

You obviously haven't been working with companies in the financial sector. Those Fortran developers you always hear about, this is were they are working. Many companies in that sector use rather old systems that can't be replaced. Some of those systems rely on IE6 or IE7.

I do work for a company that targets that sector and we calculate with a user base of about 50% that still use one of those browsers. Yet, we want them to have all the bells and whistles that the web offers today as well. This will change (and is changing) but rather slowly. Some might even say it barely even moves.

This is also the case in the Pharmaceutical sector. The IE6 user base accounts for the majority of the users and therefor it was the primary development platform but still requiring the bells and whistles.

And for the utility sector in they Bay Area. Legacy systems were actually built for IE6. It's so proprietary and ingrained into the work processes, that in order to use current versions of vendor products they have to install Firefox! Ha!

Perhaps because they are an important source of customers? I agree with the sentiment of pushing things further along, but the cost/benefit tradeoffs can be radically different depending on individual business requirements.

People at several online retail sites that I've spoken with say that IE6/7 clicks and conversions are proportionally better than other browsers. Nobody seems to know the reason. Maybe people are buying stuff from crappy corporate XP systems on their lunch hours, or perhaps it's computer-illiterate octogenarians spending their Social Security checks.

As another sorta-related data point, I do know that back in the late 1990s the AOL users also clicked and converted at a higher rate than the Internet in general.

> If corporations are too risk-averse to get rid of a clearly-broken tool, why should we help them out?

Because we want their business. There's a simple equation here- if you don't support IE6, you lose business from people that use it. For some web sites (such as mine, fortunately) IE6 and IE7 usage is so low that it's a rounding error. A lot of other sites are not so fortunate.

It's unfair to assert that not making things look the same (like using something e.g. CSS PIE) means we're not supporting these browsers.

Making your site work in IE6 is one thing... making the experience nice and fast is another... but it's impractical to make sure the pixels that you see in IE6 match the pixels you see in Chrome.

> There's a simple equation here- if you don't support IE6, you lose business from people that use it.

The equation is not so simple after all. And it means that you're responsible for delivering a good UX, not antiquated notions of identical rendering.

> But as we all know, due to Internet Explorer's lack of support for any of these features, we must be patient and refrain from using them, and make do with the same old tedious techniques for the foreseeable future.

Couldn't disagree with this more. Let your site have square corners, no shadows, and no animations in IE. The users won't know or care.

This. Your documents should use progressive enhancements that make them at least accessible (if not pretty) with every browser that has ever existed. If you write something that only works on FF5 or IE9, you did something wrong, and someday we'll be cursing you just as you curse people who wrote stuff that only work on IE6.

Even though I agree - this would be VERY helpful in a situation where a manager doesn't understand that it's ok for a site to look different between browsers.

I've had times where I've explained how a site will look different in older browsers, had that approved and build the site, but when it was time to look at the final product one person cant understand/accept that buttons "don't look 3D" in their browser of choice.

Not this. My job as a web developer is to create a great experience for the users of my sites, which often includes beautiful design. Saying "the users won't know or care" comes off as very arrogant -- how do you know that?

There is a branch of PIE that supports text-shadow which makes it even more useful.

I used it on my site re-design because we have multiple text-shadows creating a graphical effect on our logo without images.

It works really well and performance doesn't suffer noticeably. In fact with multiple box-shadows IE8 was faster to scroll than Safari 5.0!

If the features of this library continue to expand I can see this being a very powerful tool. As lot of other commentators have pointed out despite having so many modern browser solutions a lot of large corporate environments are slow to update. Quite often this is because they have large line-of-business apps that are written to target certain browsers (especially if ActiveX is heavily used) as well as having a culture where even a version bump in what web browser is used takes a committee and months of approval.

I ran into this problem years ago when I was interning at one of the major auto manufactures. I needed to provide web-based charting but was unable to get approval to use an open source toolkit. I ended up having to write my own library and did so using SVG. It worked no problem on the Solaris workstations running old-school Mozilla (even still had the old Netscape icon) but for the PCs running IE6 I had to get approval to install the Adobe plugin.

I use PIE for a charting software we created at my company. Here are some problems we had:

The latest release of PIE is used for IE7 and IE8 but we had some issues with it in IE6 where it would throw JavaScript errors. In addition, there is a problem with hidden elements (style="display: none") for example PNGs with transparent backgrounds or elements that have a linear background and rounded corners. If those become visible without reloading the page, the styles might not be applied correctly. You may find black images instead of the PNGs or missing element styles. This cannot be reproduced regularly, though. I found a beta version somewhere in the message boards that fixes the issues but does not work well with other versions of IE.

We never had any performance issues although we quite heavily use layout elements from CSS3.

BTW: The main website can be reached here http://css3pie.com

Anyone have experience using things like this? I've just now heard about these behaviors and I worry about performance..

I used it on a recent website project. It worked great for small forms (3-5 fields), which needed rounded corners and drop-shadows on the fields. It did not work so well for our dropdown navigation menu, things just seemed to be complicated when combining all the show/hide/hover css for that with the PIE effects. (So we wound up using traditional non-semantic markup for rounded/shadowed background images on the menus, but kept PIE for the more straightforward form fields).

Others have reported problems of slowness when used on too many elements in a page, but we never experienced this (because we didn't have "too many" elements on the page).

If you're already familiar with how to make IE-compatible shadows/corners by faking it with background images, I'd probably just stick with that. But if you want to forge ahead with CSS3 but also have an easy drop-in solution that gets you 80% there on IE, PIE is a good choice.

that is allot of code (== effort) for making things looks good (aesthetically) on IE 6-8.

I see there files created few days ago, and I wonder, why in July 2011 people and corporate will stilll choose to use IE6/7/8 when FF/Chrome/Safari available for Windows, for free, and support all or most advanced, neat HTML5 features

Should not corporate care most about information security? How can they let instances of those versions of IE in an organization's network?

The problem for corporate environments is that they tend to have a few really big legacy applications which only function correctly in older versions of IE.

The question then becomes whether to upgrade the application (with a massive price tag), or keep their old version of IE as the standard (which appears to cost nothing). In fact, there are costs of maintenance and security, but those aren't readily apparent on the balance sheets.

That is the case at my company, over half of the web based applications that we have to use are built exclusively for IE. It is a massive pain, but I do not see it changing any time soon.

Two factors. One is laziness. Upgrading to Firefox and Chrome takes work. Second is that if they upgraded, there are probably a lot of people who would complain, "WTF where is my [some menu item from IE]?"

These are stupid reasons, but they are enough to prevent companies from upgrading.

There is also the cost of labor associated with upgrading. A lot of corporate IT departments don't want to spend the hours upgrading their whole infrastructure when they have other things to do - from my experience IT is expected to run with as little cost as possible - if everything needed for people to complete their work is possible with ie6, there is no incentive to upgrade.

This, a thousand times this. IE6 persists in corporations not so much from risk-aversion, but simply from inertia and the lack of a cost-benefit justification to upgrade. Windows XP is the real culprit. XP installers even through the latest service packs never came with anything newer than IE6, so it's going to stick around as long as XP does. And given the frequency that most corporate workstations get reassigned and reinstalled and reimaged, IE6 comes out of the reinstallation casket all the time.

And consider the users as well. These are corporations with thousands of users with no technical expertise, especially corps in fundamentally non-technological fields like say education or food service. These are the users with twelve handwritten steps on Post-its for how to open their timesheet on the intraweb. Even a switch to tabbed browsing can get them lost in their routines. What IT department wants to spend any amount of time upgrading and retraining these users?

"twelve handwritten steps on Post-its for how to open their timesheet on the intraweb" It's funny to watch this in action.

As others have stated, there are a ton of web apps that require IE. And once you have to install IE for them, why install others? So that your employees can surf all the web sites they want on company time? The inventive to provide that is not very high.

Is the name a homage to PositionIsEverything.net? The quintessential IE6 CSS bug documenter since the first wave of web standards?

You know what's a progressive Internet Explorer? IE 9 or 10.

Let's quit trying to duct tape the web for IE's sake.

Just the tool I was looking for, but the current support for CSS3 rules is very limited, I hope they add more implementations for IE. Thanks Jason.

It's working for me. There's about a 1 second delay before the CSS3 styles kick-in. But that's not a deal-breaker.

This is fantastic.

For what it's worth, we used PIE for IE7 support of some css3 features at my last job. Performance was awful and buggy. I do not recommend trying to force non-essential graphical niceties onto browsers that can't support them. If you or your boss is insisting on some insane pixel-per-pixel matching between all browsers, either you or they are wrong.

Reply on news.ycombinator.com